HomeArticle

The founding member of the former Baidu driverless vehicle team joins Changan, and the "end-to-end" intelligent driving will be implemented within this year in Changan. | Exclusive from 36Kr.

李安琪2025-02-14 18:37
The landing pace of Changan's intelligent driving this year may be more compact.

Written by Li Anqi

Edited by Li Qin

36Kr Auto has exclusively learned that Pan Yifeng, a former veteran of Baidu's unmanned vehicle project, has joined Changan Automobile's intelligent driving team and reports to Tao Ji, the Chief Intelligent Driving Technology Officer. According to an informed source, after joining Changan's intelligent driving team, Pan Yifeng is mainly responsible for intelligent driving algorithm software, data closed-loop, and other directions.

Pan Yifeng is a Ph.D. from the Institute of Automation, Chinese Academy of Sciences. He joined Baidu in 2013 and is a founding member of Baidu IDL (Institute of Deep Learning) and the founding member of Baidu's unmanned vehicle team. He previously served as the technical director of perception, prediction, planning and control, and vehicle-road coordination for Baidu Robotaxi (autonomous driving taxi).

In the middle of 2023, Pan Yifeng joined the intelligent driving company Jianzhi Robotics as a partner, during which he led the team to provide assisted driving software solutions for some BYD models.

The addition of Pan Yifeng is a strong evidence of Changan Automobile's acceleration in intelligent driving.

On February 9 this year, Changan announced a series of intelligentization plans at the conference: Starting this year, it will no longer develop non-intelligent new products, and it will launch 35 digital and intelligent new vehicles in the next three years. At the same time, in the dimension of intelligent driving, Tian Shu Intelligent Driving will achieve full-scenario L3-level autonomous driving in 2026, and further realize full-scenario L4-level functions in 2028.

In terms of the implementation time, Changan Intelligent Driving plans to be the first to equip laser radars on 100,000-yuan-level vehicle models in August. This is an important measure for Changan to achieve "intelligent driving equality" this year.

An informed source told 36Kr Auto that the implementation rhythm of Changan Intelligent Driving this year is closer than the nodes announced at the conference, and it is simultaneously promoting self-developed intelligent driving solutions with different configurations.

On the one hand, based on the mid-level computing power platform of Horizon, several vehicle models will implement the high-speed NOA function to promote the popularization of intelligent driving;

On the other hand, in the field of high-level intelligent driving, Changan Qiyuan E07 is the first model equipped with Changan's high-level intelligent driving, with a starting price of 199,900 yuan. This model uses the NVIDIA Orin X chip and is equipped with a laser radar. It has previously pushed the high-speed NOA function to users, and the subsequent OTA upgrade to the urban NOA function is available.

According to 36Kr Auto, Changan Intelligent Driving plans to implement the "end-to-end" high-level intelligent driving solution with urban NOA function within this year.

In response to the above information, 36Kr Auto sought verification from Changan Automobile. As of the time of publication, the official has not responded.

Currently, Changan's intelligentization adopts two routes: solution procurement and self-developed intelligent driving. In the Avita brand, a joint venture with Huawei and CATL, Changan has in-depth cooperation with Huawei ADS Intelligent Driving; in the Deepal brand, some vehicle models cooperate with Huawei Intelligent Driving, and some vehicle models are left for Changan Intelligent Driving's self-developed solution.

Changan Qiyuan has become an important testing ground for the self-developed intelligent driving team. According to 36Kr Auto, before the establishment of the self-developed team, Changan Intelligent Driving was mainly led by Changxian Technology, a joint venture between Changan and Horizon.

In the second half of 2023, after Tao Ji, a former expert in Baidu's unmanned vehicle project, joined, a self-developed automotive team of over 1,000 people has been established. Previously, Han Junyu, the former head of Baidu's Visual Technology Department, has also joined the Changan Intelligent Driving team, mainly responsible for the research and development of large models and multi-modal models.

After Pan Yifeng joined and led the fields of intelligent driving algorithms, it is expected to further accelerate the speed of Changan's self-developed intelligent driving. An industry veteran told 36Kr Auto that Pan Yifeng's technical stack is very comprehensive, and he has directly worked on intelligent driving algorithms from perception to planning and control, from rules to end-to-end, and from L4 to mass production.

Zhu Huarong, the chairman of Changan Automobile, has publicly stated that Changan Automobile will cumulatively invest 200 billion yuan by 2030 and add a scientific and technological innovation team of more than 10,000 people. According to 36Kr Auto, Changan Intelligent Driving will continue to expand the team this year and is currently recruiting mass production experience talents from the first-tier forces.

Compared with the less than 10% penetration rate of high-level intelligent driving last year, the sinking and popularization of high-level intelligent driving by automakers this year will be quite fierce, basically calculated by the million-unit carrying capacity. The product scale, cost control, and iteration speed of the automakers' intelligent driving teams are all crucial in the battlefield competition.

Almost at the same time as Changan, BYD has launched the intelligent driving equality campaign for the first batch of 21 vehicle models, and even 70,000-yuan-level vehicles can have high-level intelligent driving functions. This is a gamble that can only move forward and cannot fall behind. The faster the pace and the quicker the adjustment, the more likely it is to become the epicenter and confront strong rivals.